issue 59: mandating character encoding - seeking clarification

During the F2F there was some discussion around this issue. The initial
issue text was about whether or not the working group would possibly
mandate a character set for SOAP messages. Since that time, the infoset
approach has been adopted and there is a strong case for saying that
wire encodings (including character encodings) are features of a
particular binding. This does not make the issue evaporate, however,
we just need to reword it to say:

"Should the normative HTTP binding mandate a particular character
encoding and if so, what should it be?"

This raises a number of relevant points for discussion:

 . should a binding have a 'default' encoding and scope to support other
   encodings?

 . is an an HTTP binding that specifies UTF8 the same binding as one that
specifies UTF16?

 . UTF8 and UTF16 is not enough information, there is dialect information
required to
   know what character set is applied to the encoding, how can this be
enabled?

The latter point has led the participants to seek clarification of XML spec
with
respect to UTF8/16 encoding of UCS multi-octet character sets.

 regards
   --oh

Received on Wednesday, 12 December 2001 12:49:56 UTC